Limits of Scalar Descriptors in Chiral Magnet Screening

Scalar magnetic descriptors are useful only when omitted information does not change the selected response beyond the required accuracy. We combine a literature audit with one-dimensional and three-dimensional tests. In an eight-material historical ledger, room-temperature labels coincide with magnetic-order eligibility, preventing identification of an additional descriptor effect. A published boundary benchmark resolves size dependence at fixed material descriptor. In a finite chiral-spin model with complete dipolar interactions, four near-matched-radius pairs at the same initial field and pairwise equal thickness have different field-step responses. First-order source-state susceptibility fails its declared twofold-improvement criterion. A disclosed second-order extension, derived from equilibrium response, is tested on eight new settings. Its root-mean-square radius error is 0.000259 lattice spacings, compared with 0.006649 for the best tested scalar normalization and 0.002221 for a multivariable baseline. Independent dipole calculations, stationarity refinement, contour checks and isolated prediction replay establish the numerical scope. The method requires the complete source configuration and Hamiltonian derivatives. The results demonstrate a target-specific audit and mechanistic restoration within declared finite geometries; they do not establish a low-cost material-only descriptor, calibrated lifetime or universal material optimum.